1. Field of the Invention
The present invention relates to a phase-change memory device and method of manufacturing the same, and more particularly to a phase-change memory device and method of manufacturing the same, which can reduce a contact area between a bottom electrode and a phase-change layer, thereby reducing quantity of current.
2. Description of the Prior Art
Recently, increasing attentions are paid to flash memory devices having a capability of keeping information stored therein even when power supplied thereto is interrupted. Usually, a flash memory includes a tunnel oxide layer, a floating gate, an inter-gate dielectric layer, and a control gate electrode, which are sequentially laminated on a channel. Therefore, in order to improve the reliability and program efficiency of the flash memory devices, it is necessary to improve the film quality of the tunnel oxide layer and increase a coupling rate of each cell.
Further, in place of the flash memory devices, new non-volatile memory devices have recently proposed. For example, a phase-change memory device stores information by using electric resistance difference according to the phase change. In such a phase-change memory device, a thin film made from chalcogenide alloy (Ge2Sb2Te5) experiences a phase change from an amorphous state to a crystalline state, so that the resistance and activation energy of the chalcogenide layer decrease while both the long-range atomic order and free electron density increase. The phase-change memory device is advantageous in that the phase-change memory device can be easily fabricated as an SOC (System On Chip) and is relatively low priced. Further, the phase-change memory device has a very fast processing speed of 5 ns, consumes a reduced power, and has a wide operation temperature range of −196˜180° C.
FIG. 1 is a graph for illustrating a method of programming and erasing a phase-change memory cell.
As shown in FIG. 1, if a phase-change layer is cooled after being heated at a temperature higher than the melting temperature Tm during a first operation time interval T1 (as shown by curve A), the phase-change layer comes into an amorphous state. In contrast, if the phase-change layer is cooled after being heated at a temperature lower than the melting temperature Tm and higher than the crystallization temperature Tc during a second operation time interval T2 longer than the first operation time interval T1 (as shown by curve B), the phase-change layer comes into a crystalline state. Here, the phase-change layer has a higher specific resistance in the amorphous state than in the crystalline state.
Therefore, whether the information stored in the phase-change memory cell refers to a logic ‘1’ or ‘0’ can be determined by detecting the current flowing through the phase-change layer in the ‘read’ mode. A compound material layer (hereinafter, referred to as “GST layer”) consisting of germanium (Ge), stibium (Sb), and tellurium (Te) is widely used as the phase-change layer.
FIG. 2 is a sectional view of a conventional phase-change memory cell.
In the conventional phase-change memory cell as shown in FIG. 2, a dielectric interlayer 5 is formed on a bottom electrode 3 of a semiconductor substrate 1. Then, the dielectric interlayer 5 is etched to form a contact plug 7 electrically connected with source regions and a phase-change layer 9 is then formed on the resultant substrate having the contact plug 7 formed thereon. Thereafter, a top electrode 11 is formed on the phase-change layer 9.
When voltage is applied in order to program the phase-change memory cell, heat is generated at an interface between the phase-change layer 9 and the contact plug 7, so that a portion 9a of the phase-change layer 9 is transformed into an amorphous state. The heat generated at the periphery C of the contact plug 7 and the phase-change layer 9 may be spread to the neighboring contact plug 7 and be unable to reach the temperature necessary for the phase change. Therefore, even after the phase-change layer is phase-changed into an amorphous state, there may remain an abnormal region (which is not changed into the amorphous state) at the periphery of the phase-change layer 9.
Further, in the ‘read’ and ‘write’ operations of the phase-change memory device, the large contact area between the bottom electrode and the phase-change layer requires an increased quantity of current for the phase change, thereby having a bad influence on the speed of the phase-change memory device.